Commits · 284cc922757b0494a5920f70d4e31feeb4bdfea0 · OpenDAS / vllm_cscc

24 Oct, 2025 5 commits

[MISC] `cudagraph_capture_sizes` related improvements (#26016) · 284cc922

fhl2000 authored Oct 24, 2025

Signed-off-by: fhl <2410591650@qq.com>
Signed-off-by: fhl2000 <63384265+fhl2000@users.noreply.github.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

284cc922

[MM][Bugfix] Replace `PatchEmbed`'s conv3d to linear layer (#27418) · 42efe609
Isotr0py authored Oct 24, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Roger Wang <hey@rogerw.io>
```
42efe609
[Kernel] Add GPTQv2 format support for low-bit or asymmetric quantization, by... · 5cc6bddb
Xiangyu Li authored Oct 24, 2025
```
[Kernel] Add GPTQv2 format support for low-bit or asymmetric quantization, by adapting gptq_gemm (#26092)
```
5cc6bddb
Fix pooling adapters for Transformers backend (#27338) · 1f9460c4
Harry Mellor authored Oct 24, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
1f9460c4

Granite 4.0 quark quantization support (#26944) · 70022ffc

xiao-llm authored Oct 23, 2025


Signed-off-by: Xiao YU <Xiao.YU@xilinx.com>
Signed-off-by: Xiao Yu <xiao.yu.dc@outlook.com>
Co-authored-by: Xiao YU <Xiao.YU@xilinx.com>

70022ffc

23 Oct, 2025 11 commits
- [Hardware][POWERPC] Disable oneDNN path in vllm/model_executor/layers/utils.py for Powerpc (#27422) · f417746a
  Akash kaothalkar authored Oct 24, 2025
```
Signed-off-by: Akash Kaothalkar <akash.kaothalkar@ibm.com>
Co-authored-by: Akash Kaothalkar <akash.kaothalkar@ibm.com>
```
  f417746a
- [Model] Siglip Embedding Support (#27324) · 0552cfb1
  Yu Jiaqi authored Oct 24, 2025
```
Signed-off-by: piood <2477084691@qq.com>
```
  0552cfb1
- [Chore] Separate out `vllm.utils.platform_utils.py` (#27374) · ca76486a
  Jonathan Chen authored Oct 23, 2025
```
Signed-off-by: Jonathan <chenleejonathan@gmail.com>
```
  ca76486a
- [Bugfix] Fix AWQ marlin layer skipping (#27416) · 81d5bb76
  Isotr0py authored Oct 24, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  81d5bb76
- [Bugfix][ROCm][DeepSeek] Fix for forward_hip in rope for DeepSeek (#27373) · 0825197b
  Gregory Shtrasberg authored Oct 23, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
```
  0825197b
- [Bugfix] Fix dp_chunking enablement logic in FusedMoE layer (#27220) · 9ef3d5b8
  Alexander Matveev authored Oct 23, 2025
```
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
```
  9ef3d5b8
- [Frontend][4/N] Improve all pooling task | Add plugin pooling task (#26973) · 3fa2c121
  wang.yuqi authored Oct 23, 2025
```
Signed-off-by: wang.yuqi <noooop@126.com>
Signed-off-by: Christian Pinto <christian.pinto@ibm.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Christian Pinto <christian.pinto@ibm.com>
```
  3fa2c121
- [CI/Build] Remove unnecessary flags from test registry (#27353) · fe2016de
  Cyrus Leung authored Oct 23, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  fe2016de
- [Bugfix] Honor --mm_encoder_attn_backend when used (#27124) · 570c3e1c
  Bradley D authored Oct 23, 2025
```
Co-authored-by: Bradley D <4551889+bradleyhd@users.noreply.github.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
```
  570c3e1c
- [Model] Add MoE support for NemotronH (#25863) · 61089465
  tomeras91 authored Oct 23, 2025
```
Signed-off-by: Tomer Asida <57313761+tomeras91@users.noreply.github.com>
```
  61089465
- [Bugfix] Fix deepseek-ocr multi-image inference and add... · 2566dca2
  Isotr0py authored Oct 23, 2025
```
[Bugfix] Fix deepseek-ocr multi-image inference and add `merge_by_field_config=True` with tensor schema support (#27361)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  2566dca2
22 Oct, 2025 8 commits
- [Model] Revert PR #26715: Restore custom PaliGemma and Gemma3-MM impl… (#27309) · e05a6754
  Luciano Martins authored Oct 22, 2025
```
Signed-off-by: Luciano Martins <lucianommartins@users.noreply.github.com>
Co-authored-by: Luciano Martins <lucianommartins@users.noreply.github.com>
```
  e05a6754
- [Bugfix] Fix HF format InternVL large variants video processing (#27330) · db6f28d8
  Isotr0py authored Oct 22, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  db6f28d8
- [Bugfix] Make `get_mrope_input_positions` instance methods (#27342) · 14e2f123
  Cyrus Leung authored Oct 22, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  14e2f123
- [bugfix] remove unused parameters to reduce unnecessary vram usage (#26789) · 980de31c
  Reinforce-II authored Oct 22, 2025
```
Signed-off-by: Reinforce-II <fate@eastal.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
```
  980de31c
- [Bug] Fix DeepSeek-V2.5-1210-FP8 issue (#27267) · 1c160841
  Wentao Ye authored Oct 22, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  1c160841
- [Model] Upstream Deepseek-OCR model (#27247) · 675aa2ec
  Isotr0py authored Oct 22, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
```
  675aa2ec
- [Chore] Separate out optional dependency checks from vllm.utils (#27207) · 3ae082c3
  dongbo910220 authored Oct 22, 2025
```
Signed-off-by: dongbo910220 <1275604947@qq.com>
Signed-off-by: dongbo910220 <32610838+dongbo910220@users.noreply.github.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  3ae082c3
- [torch.compile] Enable silu_mul_fp8_quant fusion without custom ops enabled (#27146) · ab3e8004
  Jiangyun Zhu authored Oct 22, 2025
```
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
```
  ab3e8004
21 Oct, 2025 14 commits

[Deepseek v3.2] Remove extra logics in indexer (#26465) · 09a7e6f6

Lain authored Oct 21, 2025


Signed-off-by: Siyuan Fu <siyuanf@nvidia.com>
Signed-off-by: Daniel Campora <961215+dcampora@users.noreply.github.com>
Signed-off-by: Lain <siyuanf@nvidia.com>
Co-authored-by: Daniel Campora <961215+dcampora@users.noreply.github.com>

09a7e6f6

[Performance] Dual stream execution of "shared_experts" and "selected_experts"... · 344a0017

Alexander Matveev authored Oct 21, 2025


[Performance] Dual stream execution of "shared_experts" and "selected_experts" inside FusedMoE (#26440)
Signed-off-by: Alexander Matveev <amatveev@redhat.com>

344a0017

[Feature] Batch Invariant for R1 TP 8 on Blackwell (#27229) · 86ed7702
Wentao Ye authored Oct 21, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
86ed7702
[ROCM] Enable CompressedTensorsWNA16 (#27187) · ba09652d
JartX authored Oct 21, 2025
```
Signed-off-by: JartX <sagformas@epdcenter.es>
```
ba09652d
[Deepseek v3.2] Optimize top_k_per_row (#26763) · 80e94529
Daniel Cámpora authored Oct 21, 2025
```
Signed-off-by: Daniel Campora <961215+dcampora@users.noreply.github.com>
```
80e94529
[MM][Core] Decouple ViT backend from LM backend (#27061) · c3a2c6ac
Roger Wang authored Oct 21, 2025
```
Signed-off-by: Roger Wang <hey@rogerw.io>
```
c3a2c6ac
[Fix][Spec Decode] Fix llama4 draft loading with different quantization (#27136) · be444507
Zebing Lin authored Oct 21, 2025
```
Signed-off-by: linzebing <linzebing1995@gmail.com>
```
be444507
[Bugfix] Fix broken MTP weight loading for FP8 KV Scales (#27227) · f381cf23
Benjamin Chislett authored Oct 21, 2025
```
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
```
f381cf23

[Bugfix] Fix gpt-oss w4a8 DP/EP on B200 (#26729) · 5ff5d94e

Varun Sundar Rabindranath authored Oct 21, 2025


Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

5ff5d94e

[ModelOpt] Load w13/w2_input_scale for all experts, nvfp4 (#26135) · f95da13c

Shu Wang authored Oct 21, 2025


Signed-off-by: Shu Wang <shuw@nvidia.com>
Signed-off-by: Shu Wang. <shuw@nvidia.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

f95da13c

[BugFix] GPT-OSS Attention DP + MoE TP weight loading issue (#24032) · aef368aa
Po-Han Huang (NVIDIA) authored Oct 21, 2025
```
Signed-off-by: Po-Han Huang <pohanh@nvidia.com>
```
aef368aa

[Feature][Kernel]FusedMoE LoRA (#21229) · 5f6cbf60

Chen Wu authored Oct 21, 2025


Signed-off-by: wuchen <cntryroa@gmail.com>
Signed-off-by: banjuede <lmklhc@163.com>
Signed-off-by: Chen Wu <cntryroa@gmail.com>
Signed-off-by: Danielle Robinson <dmmaddix@amazon.com>
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Signed-off-by: bk-201 <joy25810@foxmail.com>
Co-authored-by: wuchen <wuchen@zetyun.com>
Co-authored-by: Nathan Van Gheem <vangheem@gmail.com>
Co-authored-by: banjuede <lmklhc@163.com>
Co-authored-by: Danielle Robinson <dmmaddix@amazon.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: bk-201 <joy25810@foxmail.com>

5f6cbf60

[cpu] Dispatch un-quantized linear to oneDNN/ACL by default for AArch64 (#27183) · 163965d1
Fadi Arafeh authored Oct 21, 2025
```
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
Co-authored-by: Michael Yang <Michael.Yang@arm.com>
```
163965d1
[Quantization] Automatically infer AWQ `modules_to_not_convert` field (#26909) · 352c0c8a
Isotr0py authored Oct 21, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
352c0c8a

20 Oct, 2025 2 commits
- [Feature][Quantization] auto_round support for mixed bits quantization (#23812) · 87778d5f
  Heng Guo authored Oct 21, 2025
```
Signed-off-by: n1ck-guo <heng.guo@intel.com>
Signed-off-by: Heng Guo <heng.guo@intel.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  87778d5f
- [Kernel][Model] Tune fused_moe Triton configs for Qwen3-30B A3/A3B on H100 (FP8/BF16) (#26268) · 4d0f2661
  shivampr authored Oct 20, 2025
```
Signed-off-by: Shivam <shivampr.dev@gmail.com>
```
  4d0f2661